PLANT PROTECTION
438
55 LANDTECHNIK 6/2000Gerhard Wartenberg, Potsdam-Bornim
Site-specific precision spraying
P
ractical trials of precision site-specific herbicide spraying, e.g., on cereals and maize, indicate that in more than 2/3 of trial areas the resultant reduction of herbicide is justifiable from a crop production point of view. An off line check over several years featuring:• manual assessment
• establishment of weed maps, and
• electronically controlled application gave average savings of around 30 % on re- commended herbicide amount, or 20 to 30 DM/ha in material costs [1]. Comparative tests featuring fungicide and plant growth re- gulator spraying gave similar results [2]. But such advantages are only available where there are cost-efficient systems for precision spraying.
Realistic future advances in the develop- ment of alternative identification strategies and faster detection techniques for site-spe- cific spraying in real time [4, 5, 6] will, in the near future, require new solutions, especial- ly for hydraulic control of the sprayer.
Investigations into site-specific applicati- on and control via sensors of a commercial- ly-available sprayer, e.g. for the sensing of weed density, led to first experiences of the new demands on future spraying technology.
Targets for sprayer control
The control of sprayers in site-specific ap- plications is based on target values accor- ding to the heterogeneity and divisional structure of the areas to be treated. For her- bicide application the target in the current systems is taken from weed density or the calculated yield loss caused by the weed
growth. Target values upon which variations in the application rates of plant growth regu- lator or fungicide are also crop development (plant mass) or the surface area of the crop in question [6].
The adjustment frequency, which should transpose the regulating system, is a function of the heterogeneity division of field speci- fic parameters. Thus the target of site-speci- fic spraying economic viability is also deter- mined by the precision with which applicati- on is varied according to requirement.
Heterogeneity of weed distribution Weed population variation range and effect (density, yield loss) is large and often cha- racterised by high frequency alteration pat- terns. Extreme differences of from 500 % have been determined within a few decime- tres [7]. Often, these can be traced to hus- bandry errors. Generally, the distribution of weeds is due to the local distribution of the given heterogeneity parameter following the principle of superimposed oscillations. The
‘long wave’ transition of sectors with low or high densities is continuously overlaid by
‘short wave’ changes of smaller area. These characteristics create the seed distribution potential in the soil and the site-specific dif- ferences under emergence conditions.
Investigations within a 1 m2grid empha- sised the extent of weed population variabi- lity per site. Where the area used in calcula- tions is 0.5 m2it must be remembered that this tends to further even-out original varia- bility (fig. 1).
Not much further than 10 m laterally to the main cultivation direction of the field stretch
Precision spraying of plant protec- tion substances raises new challen- ges for sprayer technology in appli- cation. In conventional systems the regulating system has only to apply an area-associated uniform dose of spray. Site-specific precision spraying requires a variable appli- cation of spray for different doses to suit mainly small areas of hete- rogeneity regarding weed density, disease or the crop itself.
Dr.-Ing. Gerhard Wartenberg is a member of the scientific staff in the Department of Technology in Crop Production, work area Applied Technology for Plant Protection, Technology and Procedural Technology in Precision Application of Herbicides at the Institute for Agricultural Engineering, Bornim e.V. (ATB), (scientific director: Prof. Dr.-Ing. J.
Zaske), Max-Eyth-Allee 100, 14469 Potsdam-Bornim;
e-mail: gwartenberg@atb-potsdam.de The investigations were able to be carried out thanks to the support and cooperation of the company Müller Electronik, Salzkotten.
Keywords
Precision agriculture, fieldsprayer, hydraulic control
Fig. 1: Example for the hetero- geneity of yield loss of the weeds populati- on into winter wheat by highly resolution grid sampling.
small areas that require the same sort of spray treatment. Further investigation results show that in the longways direction, due to the cultivation and crop care operations in the field, larger areas of a comparable stan- dard are to be expected. If this weed popula- tion reaction is applied to the conventional working widths of sprayers it shows the ne- cessity for the creation in the future of tech- nology for sectional control of spray appli- cation.
Spray reaction to controls and practical investigations
Current generation sprayers feature a bypass system.The pto driven pump delivers against the resistance of all opened nozzle cross sec- tions. When the spray boom is activated or closed a main valve reacts to the immediate pressure increase or decrease in the nozzle pipelines through opening or closing the by- pass in the sprayer tank. A regulating stretch is seen as a way of maintaining consistent area coverage by the spray. This consists of an engine-powered throttle valve and the sensor arms for the determination of driving speed and volume flow. Both parameters along with the predetermined desired value (l/ha) of the resultant pumped volume is then calculated by the job computer. The reading of the sensor arm and the transmission of the signal to the throttle valve takes place after a constant interval of time. From the adjusting speed and the and the reading time interval there results the following speed of the ac- tual value. The control loop’s dynamics de- pends upon the proportionality of both para- meters. Used because of cost grounds con- trol loops with longer reading time intervals and slow adjusting speeds. These cushioned the over-oscillations of the control process but also led to slower following speeds for the matching of the actual value and desired value.
For field tests a standard sprayer was fit- ted for regulating according to target values.
For this special software was developed which expanded the control algorithm speci- fic to the sprayer. No change was made in the
hardware. In practical operation it was shown that a satisfactory quality regarding the following of the actual value was able to be reached only through with low frequency alterations (fig. 2)
The limited velocity of the following of the adjusting parameters – effort volumes – led, where there were sudden increases or decreases in weed density, to deviations re- presenting over 5 % of the desired value.
Generally, the investigations showed on average that deviations appeared with about 60 % of the individual values lying over this threshold.
This problem was able to be solved with the alternatives:
• sensor arm readings taking place over shor- ter time intervals (software),
• equipping throttle vale with faster motor (hardware), or
• using a proportional regulator.
Summary
Practical solutions for site-specific precision plant protection in real time can bring eco- nomic and environmentally-relevant advan- tages for the farmer. The sensoric identifica- tion of weed populations will be practical in a few years. With regard to the present deve- lopment targets of sensor development for weed identification, there are differing chal- lenges, especially for the hydraulic control of the sprayer. These challenges are in the main dependant on detection principles and the arrangement of appropriate sensors on
the sprayer (table 1).
Further challenges result from practices used nowadays by farmers in conventional spraying operations. Among the main com- ponents here are:
• the successive application of different ac- tive ingredients, on different field areas however,
• the combined use of site specific and con- ventional methods (plant protection, ferti- lising) and
• the parallel application of different me- thods with different distribution patterns of the target values (same-time application of different sensors, herbicides, plant growth regulators, fungicides).
Literature
[1] Schwarz, J., G. Wartenberg andI. Ackermann:
Process-Engineering Investigations for Site- Specific Weed Control. Proceedings 2nd European Conference on Precision Agriculture 11./15.7. Odense, Denmark, 1999, Part 2, S. 699- 708
[2] Dammer, K. H., G. Wartenberg undR. Adamek:
Sensorgestützter teilflächenspezifischer Einsatz von Fungiziden und Halmstabilisatoren in Getreide. Gesunde Pflanzen 52 (2000), H. 4, S. 105-109
[3] Biller, R. H.: Pflanzenunterscheidung und gezielter Einsatz von Herbiziden. Forschungs Report, Ztschr. der Bundesforschungsanstalt, 1998, H. 1, S. 34-37
[4] Gerhards, R., M. Sökefeld, W. Kühbauch undG.
Kifferle: Pflanzenschutz für das nächste Jahrtau- send. Noch mehr Präzision in der Unkrautkon- trolle. DLZ Sonderheft 10, 2. Auflage, ((Jahr ?)) S 35-37.
[5] Wartenberg, G.andK.-H. Dammer: Teilflächenspezi- fische Unkrauterkennung im Echtzeitbetrieb. VDI Berichte, VDI Verlag, (2000) ISBN 3-18-1449-1, (im Druck)
[6] Ehlert, D., andV. Hammen: Acquisition of Biomass in Plant Populations for site specific Farming.
International Conference on Agricultural Engineering (AgEng), 24.-27.8. 1998, Teil 2, Oslo, pp. 621-622
[7] Wartenberg, G.: Untersuchungen der Vorausset- zungen für den teilflächenspezifischen chemi- schen Pflanzenschutz. Forschungsbericht des ATB, Nr. 1996/1, 1996, 72 S.
55 LANDTECHNIK 6/2000
439
System solution Requirements
for sprayer control Sensor attached to sprayer tractor front and averaging • short adjustment times 1 over an as long as possible length of the distance • high adjusting dynamic
between sensor and spray pipeline • no over-oscillation Several sensors fitted on sprayer tractor front on • see point under 1 2 the support struts (partial working width) and • sectional control
averaging over an as long as possible length of the distance between sensor and sprayer pipeline
Sensors on the spray pipeline (sectional detection, • minimum adjustment time (ms) 3 individual nozzle control), highest development, individual • nozzle-associated control,
plot control almost without prior application, no averaging • retention of fogging quality Table 1: Influence of weed detection technique on the control the plant protection machinery
Fig. 2: Relati- onship beween desired value and actual value by the sensor control- led application into real time
